Jouni Malinen [Sat, 4 Nov 2017 09:27:00 +0000 (11:27 +0200)]
DPP: Add DPP_LISTEN and DPP_STOP_LISTEN to hostapd
This is an initial step in allowing non-operating channel operations
with DPP when hostapd is the Responder. For now, this is only used for
specifying role=configurator/enrollee and qr=mutual cases similarly to
the wpa_supplicant configuration for in Responder role. Request to use a
non-operating channel will be rejected.
Add the following vendor commands and their vendor attributes for
spectral scan.
1) QCA_NL80211_VENDOR_SUBCMD_SPECTRAL_SCAN_GET_CONFIG
Get current values of spectral parameters.
2) QCA_NL80211_VENDOR_SUBCMD_SPECTRAL_SCAN_GET_DIAG_STATS
Get stats for spectral scan debug.
3) QCA_NL80211_VENDOR_SUBCMD_SPECTRAL_SCAN_GET_CAP_INFO
Get the spectral hardware capability.
4) QCA_NL80211_VENDOR_SUBCMD_SPECTRAL_SCAN_GET_STATUS
Get the current status of spectral scan.
Sunil Dutt [Mon, 9 Oct 2017 14:32:20 +0000 (20:02 +0530)]
P2P: Continue P2P_WAIT_PEER_(IDLE/CONNECT) sequence on a listen cancel
Some drivers may accept the remain-on-channel command, but instead of
indicating start event for remain-on-channel, just indicate that the
operation has been canceled immediately. This listen cancel from the
WAIT_PEER_CONNECT state ended up in discontinuation of further
WAIT_PEER_IDLE/WAIT_PEER_CONNECT state transitions. Hence, delay the
subsequent IDLE state by 100 ms.
This commit introduces the ACS functionality for the autonomous GO. The
optional parameter <freq> in p2p_group_add is enhanced to carry a value
"acs" with the intention to select the channels among any supported
band. freq = 2 / 5 carry the need to select the channels only in the
respective bands 2.4 / 5 GHz. This functionality is on top of the host
driver's capability to offload ACS, which is advertized through
WPA_DRIVER_FLAGS_ACS_OFFLOAD.
Add QCA_WLAN_VENDOR_ATTR_LL_STATS_EXT_REPORT_TIME
and QCA_WLAN_VENDOR_ATTR_LL_STATS_EXT_MEASUREMENT_TIME to record
timestamp and duration for the last MAC counters. Some user layer
application is delay sensitive. It needs to know time stamp and
measurment duration for the counters.
Jouni Malinen [Fri, 3 Nov 2017 17:24:41 +0000 (19:24 +0200)]
DPP: Move PKEX z derivation on Responder to earlier phase
K and z can be derived already based on information available at the
time the PKEX Exchange Request is being processed, so move these there
from the PKEX Commit-Reveal Request processing since that matches the
DPP tech spec description close and allows PKEX exchange to be aborted
earlier if anything unexpected happens.
Jouni Malinen [Fri, 3 Nov 2017 15:49:45 +0000 (17:49 +0200)]
DPP: PKEX counter t
Add limit on number of failed attempts that could have used PKEX code.
If the limit (5) is reached, drop the PKEX state (including the code)
and report this on the control interface to indicate that a new code
needs to be entered due to possible attack.
Jouni Malinen [Fri, 3 Nov 2017 15:13:17 +0000 (17:13 +0200)]
DPP: Terminate PKEX exchange on detection of a mismatching code
Clean up the pending PKEX exchange if Commit-Reveal Request processing
indicates a mismatch in the PKEX code. Previously, the this case was
silently ignored and the session was left in pending state that
prevented new PKEX exchanges from getting initated. Now, a new attempt
is allowed to be initiated.
Jouni Malinen [Fri, 3 Nov 2017 14:43:58 +0000 (16:43 +0200)]
DPP: PKEX and STATUS_BAD_GROUP
Report mismatching finite cyclic group with PKEX Exchange Response using
STATUS_BAD_GROUP and provide more detailed error report over the control
interface on the peer device when this happens.
Jouni Malinen [Fri, 3 Nov 2017 08:49:45 +0000 (10:49 +0200)]
Fix Status Code in TKIP countermeasures case
The previously used WLAN_REASON_MICHAEL_MIC_FAILURE (14) value as a
response to Authentication frame or (Re)Association Request frame is not
correct since the resp value is encoded in the Status Code (not Reason
Code) field. Status Code 14 is WLAN_STATUS_UNKNOWN_AUTH_TRANSACTION
which is really what this value would have meant in the response frames.
There is no Michael MIC failure status code, so have to use the generic
"Unspecified failure" (1) reason code for these cases.
Jouni Malinen [Thu, 2 Nov 2017 19:28:58 +0000 (21:28 +0200)]
DPP: Move PKEX Commit-Reveal Response building to a helper function
This cleans up dpp_pkex_rx_commit_reveal_req() a bit and makes it easier
to add protocol testing functionality to PKEX exchange similarly to the
previously added DPP Authentication case.
Jouni Malinen [Thu, 2 Nov 2017 19:21:55 +0000 (21:21 +0200)]
DPP: Move PKEX Commit-Reveal Request building to a helper function
This cleans up dpp_pkex_rx_exchange_resp() a bit and makes it easier to
add protocol testing functionality to PKEX exchange similarly to the
previously added DPP Authentication case.
Jouni Malinen [Thu, 2 Nov 2017 19:13:43 +0000 (21:13 +0200)]
DPP: Move PKEX Exchange Response building to a helper function
This cleans up dpp_pkex_rx_exchange_req() a bit and makes it easier to
add protocol testing functionality to PKEX exchange similarly to the
previously added DPP Authentication case.
Jouni Malinen [Thu, 2 Nov 2017 10:21:00 +0000 (12:21 +0200)]
DPP: Report possible PKEX code mismatch in control interface
Indicate to upper layers if PKEX Commit-Reveal Request frame AES-SIV
decryption fails. That is a likely sign of the PKEX code mismatch
between the devices.
Jouni Malinen [Wed, 1 Nov 2017 15:14:30 +0000 (17:14 +0200)]
DPP: Enable PMF when adding wpa_supplicant network profile
DPP AKM should really require PMF to be used, but since that is not yet
explicitly required in the specification, make PMF enabled for now. For
legacy PSK cases, configure PMF to be enabled as well to support both
APs in no-PMF, optional-PMF, and required-PMF configuration.
Lior David [Wed, 1 Nov 2017 09:10:47 +0000 (11:10 +0200)]
Fix test build breakage when not compiling with ieee80211w support
Build breakage was introduced by commit d8afdb210e379f414307778b9711938fbba67f9b ('Allow EAPOL-Key messages 1/4
and 3/4 to be retransmitted for testing') for some
CONFIG_TESTING_OPTIONS=y builds without CONFIG_IEEE80211W=y.
Signed-off-by: Lior David <qca_liord@qca.qualcomm.com>
hostapd: Add Min/Max Transmit Power Capability into STA command
This provides access to the Minimum/Maximum Transmit Power Capabilitie
fileds (the nominal minimum/maximum transmit power with which the STA
is capable of transmitting in the current channel; signed integer in
units of decibels relative to 1 mW).
Signed-off-by: bhagavathi perumal s <bperumal@qti.qualcomm.com>
Check hostapd current_mode before dereferencing it in additional places
While most places using this should be for cases where the hw_features
functionality is required, there seem to be some paths that are getting
exposed in new OWE related operations where that might not be the case.
Add explicit NULL pointer checks to avoid dereferencing the pointer if
it is not set when operating with driver wrappers that do not provide
sufficient information.
Jouni Malinen [Mon, 30 Oct 2017 10:08:19 +0000 (12:08 +0200)]
tests: ap_wpa2_eap_tls_versions to test TLSv1.2 with OpenSSL 1.1
Change the test condition from "is OpenSSL 1.0.2" to "is not OpenSSL
1.0.1", so that the TLSv1.2 test step gets executed with OpenSSL 1.0.2
and 1.1 (and newer).
Jouni Malinen [Sun, 29 Oct 2017 15:13:54 +0000 (17:13 +0200)]
AP-side workaround for WNM-Sleep Mode GTK/IGTK reinstallation issues
Normally, WNM-Sleep Mode exit with management frame protection
negotiated would result in the current GTK/IGTK getting added into the
WNM-Sleep Mode Response frame. Some station implementations may have a
vulnerability that results in GTK/IGTK reinstallation based on this
frame being replayed. Add a new hostapd configuration parameter that can
be used to disable that behavior and use EAPOL-Key frames for GTK/IGTK
update instead. This would likely be only used with
wpa_disable_eapol_key_retries=1 that enables a workaround for similar
issues with EAPOL-Key. This is related to station side vulnerabilities
CVE-2017-13087 and CVE-2017-13088. To enable this AP-side workaround,
set wnm_sleep_mode_no_keys=1.
Johannes Berg [Wed, 25 Oct 2017 08:06:10 +0000 (10:06 +0200)]
wpa_auth: Deplete group rekey eloop handler for strict rekeying
When strict group rekeying is in effect, every station that leaves will
cause a rekeying to happen 0.5 s after leaving. However, if a lot of
stations join/leave, the previous code could postpone this rekeying
forever, since it always re-registers the handling with a 0.5 s timeout.
Use eloop_deplete_timeout() to address that, only registering the
timeout from scratch if it wasn't pending.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Johannes Berg [Wed, 25 Oct 2017 08:26:10 +0000 (10:26 +0200)]
Allow forcing group rekeying for testing purposes
In order to test the WoWLAN GTK rekeying KRACK mitigation, add a
REKEY_GTK hostapd control interface command that can be used at certain
points of the test.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Jouni Malinen [Sun, 29 Oct 2017 14:08:02 +0000 (16:08 +0200)]
DPP: Negotiation channel change request from Initiator
Allow the Initiator to request a different channel to be used for DPP
Authentication and DPP Configuration exchanges. This commit adds support
for this in wpa_supplicant with the optional neg_freq=<freq in MHz>
parameter in DPP_AUTH_INIT.
Jouni Malinen [Sun, 29 Oct 2017 10:23:34 +0000 (12:23 +0200)]
DPP: Allow testing override values to be cleared
This allows wpa_supplicant dpp_config_obj_override,
dpp_discovery_override, and dpp_groups_override parameters to be cleared
by setting them to a zero-length value.
Jouni Malinen [Sat, 28 Oct 2017 14:40:06 +0000 (17:40 +0300)]
DPP: Process Authentication Confirm failure cases
Process Authentication Confirm with the two failure cases defined in the
spec: STATUS_NOT_COMPATIBLE and STATUS_AUTH_FAILURE. This verifies the
{R-nonce}k2 part and reports more detailed failure reason if the message
is valid.
Jouni Malinen [Mon, 23 Oct 2017 11:14:57 +0000 (14:14 +0300)]
DPP: Allow Responder to decide not to use mutual authentication
Previously, Initiator decided whether to use mutual authentication on
its own based on having own and peer bootstrapping info. This prevented
Responder from selecting not to use mutual authentication in such a
case. Fix this by allowed Initiator to fall back to non-mutual
authentication based on Responder choice if the bootstrapping mechanism
allows this (PKEX does not; it mandates use of mutual authentication).
Jouni Malinen [Sun, 22 Oct 2017 16:03:23 +0000 (19:03 +0300)]
DPP: Move Authentication Response building into a separate function
This cleans up old dpp_auth_build_resp() (now dpp_auth_build_resp_ok())
a bit by separating initialization steps for a DPP authentication
session from the code needed to build the frame. This allows
dpp_auth_build_resp_status() to share the helper function instead of
having to maintain a duplicated message construction implementation. In
addition, this makes it easier to remove some of the attributes for
protocol testing purposes.
Jouni Malinen [Sun, 22 Oct 2017 14:08:25 +0000 (17:08 +0300)]
DPP: Move Authentication Request building into a separate function
This cleans up dpp_auth_init() a bit by separating initialization steps
for a DPP authentication session from the code needed to build the
frame. In addition, this makes it easier to remove some of the
attributes for protocol testing purposes.
Jouni Malinen [Sun, 22 Oct 2017 09:35:06 +0000 (12:35 +0300)]
DPP: Report transmitted messages as control interface events
This is helpful for testing purposes and also for upper layer components
that may want to show more detailed progress through a DPP exchange.
Both the DPP-TX and DPP-TX-STATUS events are provided.
Jouni Malinen [Sun, 22 Oct 2017 08:46:12 +0000 (11:46 +0300)]
DPP: Remove unnecessary Wrapped Data checks from callers
Now that dpp_check_attrs() takes care of verifying that no attributes
are after the Wrapped Data attribute, the duplicated checks in hostapd
and wpa_supplicant side of the implementation can be removed.
Jouni Malinen [Sun, 22 Oct 2017 08:15:21 +0000 (11:15 +0300)]
DPP: Protocol testing framework
Add a generic mechanism for configuring the DPP implementation to behave
in particular different (mostly incorrect) ways for protocol testing
purposes. The new dpp_test parameter can be set to a non-zero integer to
indicate a specific behavior. This is only available in
CONFIG_TESTING_OPTIONS=y builds.
This commit include cases for an extra attribute being added after the
Wrapped Data attribute and Initiator/Responder capabilities having an
unexpected zero capability.
Jouni Malinen [Sun, 22 Oct 2017 13:42:33 +0000 (16:42 +0300)]
tests: Make sae_invalid_anti_clogging_token_req more robust
Beacon more frequently since Probe Request frames are practically ignored
in this test setup (ext_mgmt_frame_handled=1 on hostapd side) and
wpa_supplicant scans may end up getting ignored if no new results are
available due to the missing Probe Response frames.
Michael Baird [Thu, 28 Sep 2017 05:53:35 +0000 (18:53 +1300)]
EAP server: Add event messages for more EAP states
While using an external RADIUS server SUCCESS messages were not being
sent (internal was fine). Also add event messages for other states that
others might find useful, and consistency between the two.
Signed-off-by: Michael Baird <Michael.Baird@ecs.vuw.ac.nz>
projects / thirdparty / hostap.git / log
